Reaching the Other Half: Moving Towards Symptom-Based Referral Methods in a Community-Derived Sample

Background: Current research practices typically limit participation to individuals with prior clinical diagnoses. This limits findings and generalizability to families who have sought care (Durkin et al., 2015). In answer to this, approaches that do not require a previous diagnosis, as supported by Patient-Centered Outcomes Research (PCOR) health disparities methodology, reach families who have not previously accessed mental health services by sampling underserved communities referred for symptoms rather than diagnosis. This approach has the potential to broaden the generalizability of research findings and directly impact underserved children and families.

Objectives: This study seeks to understand differences between previously-diagnosed and newly-identified children with ASD and/or ADHD in a community-based sample.

Methods: Participants include 160 families referred by high-poverty schools in the Washington, DC metropolitan area for participation in a comparative effectiveness trial of two school-based executive function interventions. At this time, 94 cases have been assessed and entered. In this sample, race/ethnicity was reflective of the community (48.7% Latino, 22.7% White, 15.9% Black, 6.8% other). 27.4% of families were primarily Spanish speaking. 45.8% of the ASD group (n=24) and 52.9% of the ADHD group were newly-identified through gold standard assessment. FSIQ did not differ between children with and without prior diagnosis. At baseline, demographic information, behavioral/emotional functioning (CBCL), and family strain were measured. 

Results: Independent samples t-tests and chi-square analyses revealed no significant differences in family income and parent education level between children with and without prior diagnosis. For children with prior diagnosis, families reported significantly more strain (t(75)=-2.12, p=.037) and greater child externalizing problems (t(70)=-2.67, p=.009). There was a higher proportion of English-speaking families within children with prior diagnosis (χ²(1, N=75)=8.361, p<.006). Multiple linear regression analyses showed language spoken in the home significantly predicted prior diagnosis after accounting for behavioral/emotional problems (CBCL Total Problems; R²=.189, F(1, 67)=8.037, p=.006). Importantly, within the ASD group, neither ASD symptom severity (ADOS-2), Total Problems, family strain, nor language spoken in the home differed between those who were previously identified vs. newly diagnosed.

Conclusions: Children previously identified as ADHD were more likely to have parents who spoke English, reported feeling more parenting strain, and noticed more problems in their children. In contrast, children previously diagnosed with ASD did not differ from those newly-identified on these variables or in autism-specific symptomatology. Thus, there is value in symptom-based referral methodology for ASD studies, as these methods allow for accurate identification of individuals who meet full diagnostic criteria. Additionally, roughly half of both final diagnostic groups were newly identified. Although ADHD can be reliably diagnosed by community providers, ASD diagnosis continues to be difficult to access, and children are being missed despite showing the same degree of symptomatology. Further research should explore factors related to difficulty accessing specialized ASD diagnosis. In sum, studies allowing for symptom-based referrals, rather than diagnosis-based referrals, have the potential to impact communities and better capture the full ASD population.